From: Richard Biener <rguenther@suse.de>
To: Jakub Jelinek <jakub@redhat.com>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] Replace IFN_TRAP with BUILT_IN_UNREACHABLE_TRAP [PR107300]
Date: Thu, 2 Feb 2023 09:19:55 +0000 (UTC) [thread overview]
Message-ID: <nycvar.YFH.7.77.849.2302020919460.6551@jbgna.fhfr.qr> (raw)
In-Reply-To: <Y9t+gOQrO/wBHlxY@tucnak>
On Thu, 2 Feb 2023, Jakub Jelinek wrote:
> Hi!
>
> For PR106099 I've added IFN_TRAP as an alternative to __builtin_trap
> meant for __builtin_unreachable purposes (e.g. with -funreachable-traps
> or some sanitizers) which doesn't need vops because __builtin_unreachable
> doesn't need them either. This works in various cases, but unfortunately
> IPA likes to decide on the redirection to unreachable just by tweaking
> the cgraph edge to point to a different FUNCTION_DECL. As internal
> functions don't have a decl, this causes problems like in the following
> testcase.
>
> The following patch fixes it by removing IFN_TRAP again and replacing
> it with user inaccessible BUILT_IN_UNREACHABLE_TRAP, so that e.g.
> builtin_decl_unreachable can return it directly and we don't need to tweak
> it later in wherever we actually replace the call stmt.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK.
Thanks,
Richard.
> 2023-02-02 Jakub Jelinek <jakub@redhat.com>
>
> PR ipa/107300
> * builtins.def (BUILT_IN_UNREACHABLE_TRAP): New builtin.
> * internal-fn.def (TRAP): Remove.
> * internal-fn.cc (expand_TRAP): Remove.
> * tree.cc (build_common_builtin_nodes): Define
> BUILT_IN_UNREACHABLE_TRAP if not yet defined.
> (builtin_decl_unreachable): Use BUILT_IN_UNREACHABLE_TRAP
> instead of BUILT_IN_TRAP.
> * gimple.cc (gimple_build_builtin_unreachable): Remove
> emitting internal function for BUILT_IN_TRAP.
> * asan.cc (maybe_instrument_call): Handle BUILT_IN_UNREACHABLE_TRAP.
> * cgraph.cc (cgraph_edge::verify_corresponds_to_fndecl): Handle
> BUILT_IN_UNREACHABLE_TRAP instead of BUILT_IN_TRAP.
> * ipa-devirt.cc (possible_polymorphic_call_target_p): Handle
> BUILT_IN_UNREACHABLE_TRAP.
> * builtins.cc (expand_builtin, is_inexpensive_builtin): Likewise.
> * tree-cfg.cc (verify_gimple_call,
> pass_warn_function_return::execute): Likewise.
> * attribs.cc (decl_attributes): Don't report exclusions on
> BUILT_IN_UNREACHABLE_TRAP either.
>
> * gcc.dg/pr107300.c: New test.
>
> --- gcc/builtins.def.jj 2023-01-02 09:32:37.988059844 +0100
> +++ gcc/builtins.def 2023-02-01 19:19:15.382475912 +0100
> @@ -1048,6 +1048,7 @@ DEF_GCC_BUILTIN (BUILT_IN_SETJMP,
> DEF_EXT_LIB_BUILTIN (BUILT_IN_STRFMON, "strfmon", BT_FN_SSIZE_STRING_SIZE_CONST_STRING_VAR, ATTR_FORMAT_STRFMON_NOTHROW_3_4)
> DEF_LIB_BUILTIN (BUILT_IN_STRFTIME, "strftime", BT_FN_SIZE_STRING_SIZE_CONST_STRING_CONST_TM_PTR, ATTR_FORMAT_STRFTIME_NOTHROW_3_0)
> DEF_GCC_BUILTIN (BUILT_IN_TRAP, "trap", BT_FN_VOID, ATTR_NORETURN_NOTHROW_LEAF_COLD_LIST)
> +DEF_GCC_BUILTIN (BUILT_IN_UNREACHABLE_TRAP, "unreachable trap", BT_FN_VOID, ATTR_CONST_NORETURN_NOTHROW_LEAF_COLD_LIST)
> DEF_GCC_BUILTIN (BUILT_IN_UNREACHABLE, "unreachable", BT_FN_VOID, ATTR_CONST_NORETURN_NOTHROW_LEAF_COLD_LIST)
> DEF_GCC_BUILTIN (BUILT_IN_UNWIND_INIT, "unwind_init", BT_FN_VOID, ATTR_NULL)
> DEF_GCC_BUILTIN (BUILT_IN_UPDATE_SETJMP_BUF, "update_setjmp_buf", BT_FN_VOID_PTR, ATTR_NULL)
> --- gcc/internal-fn.def.jj 2023-01-02 09:32:38.771048530 +0100
> +++ gcc/internal-fn.def 2023-02-01 19:19:52.239936390 +0100
> @@ -457,11 +457,6 @@ DEF_INTERNAL_FN (SHUFFLEVECTOR, ECF_CONS
> /* <=> optimization. */
> DEF_INTERNAL_FN (SPACESHIP, ECF_CONST | ECF_LEAF | ECF_NOTHROW, NULL)
>
> -/* __builtin_trap created from/for __builtin_unreachable. */
> -DEF_INTERNAL_FN (TRAP, ECF_CONST | ECF_LEAF | ECF_NORETURN
> - | ECF_NOTHROW | ECF_COLD | ECF_LOOPING_CONST_OR_PURE,
> - NULL)
> -
> /* [[assume (cond)]]. */
> DEF_INTERNAL_FN (ASSUME, ECF_CONST | ECF_LEAF | ECF_NOTHROW
> | ECF_LOOPING_CONST_OR_PURE, NULL)
> --- gcc/internal-fn.cc.jj 2023-01-02 09:32:22.206287869 +0100
> +++ gcc/internal-fn.cc 2023-02-01 19:16:52.939561023 +0100
> @@ -4518,12 +4518,6 @@ expand_SPACESHIP (internal_fn, gcall *st
> }
>
> void
> -expand_TRAP (internal_fn, gcall *)
> -{
> - expand_builtin_trap ();
> -}
> -
> -void
> expand_ASSUME (internal_fn, gcall *)
> {
> }
> --- gcc/tree.cc.jj 2023-01-27 20:09:16.151971051 +0100
> +++ gcc/tree.cc 2023-02-01 20:37:01.287315982 +0100
> @@ -9758,6 +9758,7 @@ build_common_builtin_nodes (void)
>
> if (!builtin_decl_explicit_p (BUILT_IN_UNREACHABLE)
> || !builtin_decl_explicit_p (BUILT_IN_TRAP)
> + || !builtin_decl_explicit_p (BUILT_IN_UNREACHABLE_TRAP)
> || !builtin_decl_explicit_p (BUILT_IN_ABORT))
> {
> ftype = build_function_type (void_type_node, void_list_node);
> @@ -9767,6 +9768,12 @@ build_common_builtin_nodes (void)
> "__builtin_unreachable",
> ECF_NOTHROW | ECF_LEAF | ECF_NORETURN
> | ECF_CONST | ECF_COLD);
> + if (!builtin_decl_explicit_p (BUILT_IN_UNREACHABLE_TRAP))
> + local_define_builtin ("__builtin_unreachable trap", ftype,
> + BUILT_IN_UNREACHABLE_TRAP,
> + "__builtin_unreachable trap",
> + ECF_NOTHROW | ECF_LEAF | ECF_NORETURN
> + | ECF_CONST | ECF_COLD);
> if (!builtin_decl_explicit_p (BUILT_IN_ABORT))
> local_define_builtin ("__builtin_abort", ftype, BUILT_IN_ABORT,
> "abort",
> @@ -10908,7 +10915,7 @@ builtin_decl_unreachable ()
> if (sanitize_flags_p (SANITIZE_UNREACHABLE)
> ? (flag_sanitize_trap & SANITIZE_UNREACHABLE)
> : flag_unreachable_traps)
> - fncode = BUILT_IN_TRAP;
> + fncode = BUILT_IN_UNREACHABLE_TRAP;
> /* For non-trapping sanitize, we will rewrite __builtin_unreachable () later,
> in the sanopt pass. */
>
> --- gcc/gimple.cc.jj 2023-01-02 09:32:44.443966564 +0100
> +++ gcc/gimple.cc 2023-02-01 19:20:45.634154800 +0100
> @@ -430,16 +430,7 @@ gimple_build_builtin_unreachable (locati
> {
> tree data = NULL_TREE;
> tree fn = sanitize_unreachable_fn (&data, loc);
> - gcall *g;
> - if (DECL_FUNCTION_CODE (fn) != BUILT_IN_TRAP)
> - g = gimple_build_call (fn, data != NULL_TREE, data);
> - else
> - {
> - /* Instead of __builtin_trap use .TRAP, so that it doesn't
> - need vops. */
> - gcc_checking_assert (data == NULL_TREE);
> - g = gimple_build_call_internal (IFN_TRAP, 0);
> - }
> + gcall *g = gimple_build_call (fn, data != NULL_TREE, data);
> gimple_call_set_ctrl_altering (g, true);
> gimple_set_location (g, loc);
> return g;
> --- gcc/asan.cc.jj 2023-01-02 09:32:30.191172498 +0100
> +++ gcc/asan.cc 2023-02-01 20:40:02.132674664 +0100
> @@ -2951,6 +2951,7 @@ maybe_instrument_call (gimple_stmt_itera
> switch (DECL_FUNCTION_CODE (callee))
> {
> case BUILT_IN_UNREACHABLE:
> + case BUILT_IN_UNREACHABLE_TRAP:
> case BUILT_IN_TRAP:
> /* Don't instrument these. */
> return false;
> --- gcc/cgraph.cc.jj 2023-01-30 10:57:03.433236251 +0100
> +++ gcc/cgraph.cc 2023-02-01 20:37:39.878752332 +0100
> @@ -3248,11 +3248,11 @@ cgraph_edge::verify_corresponds_to_fndec
> node = node->ultimate_alias_target ();
>
> /* Optimizers can redirect unreachable calls or calls triggering undefined
> - behavior to __builtin_unreachable or __builtin_trap. */
> + behavior to __builtin_unreachable or __builtin_unreachable trap. */
>
> if (fndecl_built_in_p (callee->decl, BUILT_IN_NORMAL)
> && (DECL_FUNCTION_CODE (callee->decl) == BUILT_IN_UNREACHABLE
> - || DECL_FUNCTION_CODE (callee->decl) == BUILT_IN_TRAP))
> + || DECL_FUNCTION_CODE (callee->decl) == BUILT_IN_UNREACHABLE_TRAP))
> return false;
>
> if (callee->former_clone_of != node->decl
> --- gcc/ipa-devirt.cc.jj 2023-02-01 17:26:28.241878107 +0100
> +++ gcc/ipa-devirt.cc 2023-02-01 20:42:46.472274417 +0100
> @@ -3471,8 +3471,10 @@ possible_polymorphic_call_target_p (tree
> unsigned int i;
> bool final;
>
> - if (fndecl_built_in_p (n->decl, BUILT_IN_UNREACHABLE)
> - || fndecl_built_in_p (n->decl, BUILT_IN_TRAP))
> + if (fndecl_built_in_p (n->decl, BUILT_IN_NORMAL)
> + && (DECL_FUNCTION_CODE (n->decl) == BUILT_IN_UNREACHABLE
> + || DECL_FUNCTION_CODE (n->decl) == BUILT_IN_TRAP
> + || DECL_FUNCTION_CODE (n->decl) == BUILT_IN_UNREACHABLE_TRAP))
> return true;
>
> if (is_cxa_pure_virtual_p (n->decl))
> --- gcc/builtins.cc.jj 2023-01-02 09:32:20.735309123 +0100
> +++ gcc/builtins.cc 2023-02-01 20:39:28.727162564 +0100
> @@ -7844,6 +7844,7 @@ expand_builtin (tree exp, rtx target, rt
> break;
>
> case BUILT_IN_TRAP:
> + case BUILT_IN_UNREACHABLE_TRAP:
> expand_builtin_trap ();
> return const0_rtx;
>
> @@ -11310,6 +11311,7 @@ is_inexpensive_builtin (tree decl)
> case BUILT_IN_VA_ARG_PACK_LEN:
> case BUILT_IN_VA_COPY:
> case BUILT_IN_TRAP:
> + case BUILT_IN_UNREACHABLE_TRAP:
> case BUILT_IN_SAVEREGS:
> case BUILT_IN_POPCOUNTL:
> case BUILT_IN_POPCOUNTLL:
> --- gcc/tree-cfg.cc.jj 2023-01-02 09:32:27.879205904 +0100
> +++ gcc/tree-cfg.cc 2023-02-01 20:44:44.826545812 +0100
> @@ -3503,6 +3503,7 @@ verify_gimple_call (gcall *stmt)
> switch (DECL_FUNCTION_CODE (fndecl))
> {
> case BUILT_IN_UNREACHABLE:
> + case BUILT_IN_UNREACHABLE_TRAP:
> case BUILT_IN_TRAP:
> if (gimple_call_num_args (stmt) > 0)
> {
> @@ -9681,6 +9682,8 @@ pass_warn_function_return::execute (func
> && ((LOCATION_LOCUS (gimple_location (last))
> == BUILTINS_LOCATION
> && (gimple_call_builtin_p (last, BUILT_IN_UNREACHABLE)
> + || gimple_call_builtin_p (last,
> + BUILT_IN_UNREACHABLE_TRAP)
> || gimple_call_builtin_p (last, BUILT_IN_TRAP)))
> || gimple_call_builtin_p (last, ubsan_missing_ret)))
> {
> --- gcc/attribs.cc.jj 2023-01-13 17:37:45.056485561 +0100
> +++ gcc/attribs.cc 2023-02-01 20:54:16.134198166 +0100
> @@ -846,6 +846,7 @@ decl_attributes (tree *node, tree attrib
> || !DECL_P (*anode)
> || DECL_BUILT_IN_CLASS (*anode) != BUILT_IN_NORMAL
> || (DECL_FUNCTION_CODE (*anode) != BUILT_IN_UNREACHABLE
> + && DECL_FUNCTION_CODE (*anode) != BUILT_IN_UNREACHABLE_TRAP
> && (DECL_FUNCTION_CODE (*anode)
> != BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE)))
> {
> --- gcc/testsuite/gcc.dg/pr107300.c.jj 2023-02-01 20:55:36.602021915 +0100
> +++ gcc/testsuite/gcc.dg/pr107300.c 2023-02-01 20:55:25.298187139 +0100
> @@ -0,0 +1,19 @@
> +/* PR ipa/107300 */
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -fipa-cp-clone -funreachable-traps -fno-inline" } */
> +
> +void
> +bar (int x, int y)
> +{
> + if (x)
> + __builtin_unreachable ();
> +
> + if (y)
> + __builtin_abort ();
> +}
> +
> +void
> +foo (void)
> +{
> + bar (0, 0);
> +}
>
> Jakub
>
>
--
Richard Biener <rguenther@suse.de>
SUSE Software Solutions Germany GmbH, Frankenstrasse 146, 90461 Nuernberg,
Germany; GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman;
HRB 36809 (AG Nuernberg)
prev parent reply other threads:[~2023-02-02 9:19 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-02 9:12 Jakub Jelinek
2023-02-02 9:19 ` Richard Biener [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=nycvar.YFH.7.77.849.2302020919460.6551@jbgna.fhfr.qr \
--to=rguenther@suse.de \
--cc=gcc-patches@gcc.gnu.org \
--cc=jakub@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).